Citation

Hobday M, Meuleners L. Accid. Anal. Prev. 2018; 113: 117-124.

Affiliation

Curtin-Monash Accident Research Centre, GPO Box U1987, Perth WA 6845, Australia. Electronic address: L.meuleners@curtin.edu.au.

Copyright

(Copyright © 2018, Elsevier Publishing)

DOI

10.1016/j.aap.2018.01.013

PMID

29407658

Abstract

This study examined the effects of distance from alcohol outlets to motor vehicle crashes across the Perth metropolitan area. A retrospective population-based study was undertaken using measures of alcohol- and non-alcohol-related crashes, and their proximity to alcohol outlets, using a geographic information system. Two logistic regression models were developed with the following outcomes: i) crashes including drivers with BAC ≥ 0.05%, and ii) weekend single vehicle night-time crashes, a surrogate measure of alcohol-related crashes. The surrogate measures of non-alcohol-related crashes for these models were all day-time and single vehicle day-time crashes respectively. The major predictors of alcohol-related crashes were number of on-premise outlets and bottleshops in buffer zones up to 2 km, 2 km-5 km, 5 km-10 km and 10 km-20 km from crashes. The distance from the central business district (CBD) and sociodemographic factors were controlled for. The study included 341,467 crashes that occurred between 2005 and 2015. The highest crash incidence rates occurred in the CBD. The statistical models indicated that crashes with a higher number of on-premise outlets in adjacent buffer zones were more likely to be alcohol-related than non-alcohol-related crashes. Crashes with a higher number of on-premise outlets less than 2 km, 2 km-5 km, 5 km-10 km, and 10 km-20 km from the crashes were significantly more likely to be weekend single vehicle night-time crashes than day-time crashes (OR = 1.014; 95% CI:1.002-1.027, OR = 1.022; 95% CI:1.014-1.029, OR = 1.019; 95% CI:1.014-1.024, and OR = 1.017; 95% CI:1.014-1.020 respectively). There was some evidence that crashes with lower number of bottleshops in adjacent buffer zones were more likely to be alcohol-related crashes, although this was not consistent across both models and all buffer zones. When other predictors were controlled for, alcohol-related crashes were more likely to occur further from the CBD, than in the CBD. Recommendations about the timing and location of roadside alcohol testing are made.

Copyright © 2018 Elsevier Ltd. All rights reserved.

Language: en

Keywords

Alcohol; Alcohol availability; Crash risk; Drink-driving; Safe system approach